Unmount fuse missing

[nezter]

@lucashalbert Starting off this docker image has been the best out of all that I have tested. Non have been as easy and less errors.

This may be a 100% user error. But is there an unmount function built in? It locks the folder up if the instance goes down. Maybe adding a check on boot for mount if present unmount.

Also attaching a persistent volume in the rlcone instance the files are present on mount. But attaching the volume to another instance shows the mount folder blank. Yet creating a file on the persistent volume shows on all instance it is attached too.

[lucashalbert]

@nezter, I'm glad to hear that you like the image! Please continue to report any bugs/issues you find with it.

As for your question regarding unmounting, the recommended method is to use Ctrl-C or send the SIGINT or SIGTERM signal to the process. When using this image, you should be able to simply stop the container which in turn will send the SIGTERM signal to the container's root process (PID 1), resulting in the rclone binary being killed. I typically use this image in Kubernetes which means if, for some reason, the container dies, then the pod is destroyed and no mountpoint artifacts remain. If you do experience leftover mountpoint artifacts [ie: gcache-crypt: on /mnt type fuse.rclone (rw,nosuid,nodev,relatime,user_id=0,group_id=0,allow_other) in the output of the mount command] after the container has been stopped, try running sudo fusermount -u ${volume_mountpoint} on the Docker host. That should (in theory) clean up those leftover artifacts.

On a side note, I am planning to implement s6-overlay in this image so that proper filesystem cleanup occurs on container teardown.

Regarding the persistent volume sharing, can you be more specific as to how you are attaching the persistent volume from the rclone instance to the other running instance? My assumption is that you are running a privileged rclone container with an attached shared volume mapped to a directory on the Docker host which is then shared with another container as an attached volume. Are you using docker-compose or any other orchestration tool, or just vanilla docker?

[nezter]

@lucashalbert Look forward to the s6-overlay!
The tip on the Stop works nice. Unfortunately running on RancherOS sudo fusermount -u ${volume_mountpoint} give cannot be found on current host to unmount error.

Portainer:1.23.2
Both instances are running Privileged: true
The other two instances are,
Instance Two:
https://hub.docker.com/r/jellyfin/jellyfin
Instance Three:
https://hub.docker.com/r/linuxserver/radarr

Rclone Instance

Mounts
1
Destination /mnt
Driver local
Mode z
Name rclone
Propagation
RW true
Source /var/lib/docker/volumes/rclone/_data
Type volume
Name /rclone_mount

Secondary Instance

Mounts
0
Destination /mnt
Driver local
Mode z
Name rclone
Propagation
RW true
Source /var/lib/docker/volumes/rclone/_data
Type volume

Third Instance

Mounts
0
Destination /mnt
Driver local
Mode z
Name rclone
Propagation
RW true
Source /var/lib/docker/volumes/rclone/_data
Type volume

If I sub level the mount one directory deeper to /mnt/media/ on the volume. And I create a file on the root of the volume all three instances can see it. echo "test" >> /mnt/test_from_{$instancename}.txt.

Rclone directory of mounted Volume
/ # ls /mnt/
Output
test_from_rclone.txt test_from_instance2.txt media

/ # ls /mnt/media/
Output
mountcheck mountcheck.txt

Instance Two and Three
root@6eb9543b0a28:/# ls /mnt/
Output
test_from_rclone.txt test_from_instance2.txt media

root@6eb9543b0a28:/# ls /mnt/media/
Output
root@6eb9543b0a28:/#

[lucashalbert]

Hey @nezter, sorry that it's taken me so long to get back to you. I'm not 100% sure how to fix your problem without seeing how you have your volume mounts configured between the three containers. If you can share those configs, I may be able to assist more.

On a side note, I just published an s6-overlay version of my rclone image lucashalbert/rclone:latest-s6-beta. Please feel free to try it out and report any bugs that you encounter.

[nezter]

@lucashalbert Thank you for you reply. I was able to figure out most of the issue. It had to do with the mount commands, it happens that using Portainer removes the needed mount setting for it to properly mount on the root OS. But as far as mounting in a persistent storage device and providing access to other containers has never been solved.

Going to try out your new image soon. glad you added s6.